AB-3D-MOT, annotations & ClearMOT

https://pypi.org/project/eval-ab-3d-mot/

2.2.0

• Improve the README.
• Possibility to report on the tracker parameters to stdout in the ClavIA script.
• Set tracker parameters (algorithm, metric, threshold and max_age) via options in ClavIA CLI.
• Set tracker parameters (algorithm, metric, threshold and max_age) via options in batch run CLI.
• Possibility to report on the tracker parameters to stdout in the batch run script.
• The number of figures after floating point is 4 for F1, Precision and Recall and 6 for Accuracy.
• Report tracker parameters once, before tracking in run-ab-3d-mot-with-clavia and batch-run-ab-3d-mot.
• Script batch-run-ab-3d-mot-annotations to run the tracker consuming KITTI annotations.
• Meta info is passed through the tracker when consuming annotations.

2.1.1 Run AB3DMOT with ClavIA assessment on top of KITTI annotations

2.1.1

• Add script run-ab-3d-mot-with-clavia to run the tracker consuming KITTI annotations.
• Use pure-ab-3d-mot==2.1.0.
• Update third-party development dependencies.
• Improve the README.

2.0.0 consuming ClavIA from a package

2.0.0

• Use the AssociationQuality class from the package association-quality-clavia.

1.0.0 first version with instrumented association

1.0.0

• Added the module association_quality with an implementation of the classifier.

0.1.0 first release without ClavIA

Evaluation part of the AB3DMOT by Xinshuo Weng (https://github.com/xinshuoweng/AB3DMOT)
The purpose of the package is to enable calculation of the detection+tracking quality
metrics for 3D tracking with KITTI data set.

What's Changed

• Split evaluate module by @kovalp in #1
• Test track data and stat by @kovalp in #2
• Test tracking evaluation class by @kovalp in #3
• Test thresholds by @kovalp in #4
• Test evaluate function by @kovalp in #5
• Suggest a cli by @kovalp in #6
• Suggest a cli for running pure ab 3d mot by @kovalp in #7
• Evaluate one sequence via cli by @kovalp in #8
• Run the pure ab 3d mot for a set of detection sequences by @kovalp in #9
• Reproduce the data 0 tracking format of original ab3dmot by @kovalp in #10
• Maintenance by @kovalp in #11
• after-rerun by @kovalp in #12
• Add threshold to the batch eval by @kovalp in #13
• Add number of annotations to r cnn adaptor by @kovalp in #14
• up pure-ab-3d-mot tracker by @kovalp in #15
• Reproduce original evaluations by @kovalp in #16
• +change by @kovalp in #17

New Contributors

• @kovalp made their first contribution in #1

Full Changelog: https://github.com/kovalp/eval-ab-3d-mot/commits/0.1.0